Um sistema de apoio ao jogador para jogos de estratégia em tempo real

Abstract

Even though Real Time Strategy (RTS) games are successfully being used as a platform for Artificial Intelligence (AI) research and experimentation, existing work tends to focus only on the development ofintelligent agents (bots) for winning an RTS game. Other approaches use Machine Learning techniques to learn the rules of the game to try to predict the player's next move. These approaches tend to ignoreinteresting research topics like the development of AI agents that help the user in his games. This work presents an approach to implement an AI agent capable of helping out the player by giving him some tactical and strategical tips. The problem we are trying to solve is to improve the performance of the human RTS player, and our main objective is to develop metrics to evaluate the current game state, elaborate hypothesis on how to improve the player's and communicate this information to him formatted as a set of strategy tips. User and performance tests suggest that the experimental framework developed in this work satisfies its main objectives without degrading an existing game's computational performance.